1. Technical Field
The present invention relates to communication systems and, more particularly, analog-to-digital and digital-to-analog converters used within transceivers.
2. Related Art
Communication systems are known to support wireless and wire lined communications between wireless and/or wire lined communication devices. Such communication systems range from national and/or international cellular telephone systems to the Internet to point-to-point in-home wireless networks. Each type of communication system is constructed, and hence operates, in accordance with one or more communication standards. For instance, wireless communication systems may operate in accordance with one or more standards, including, but not limited to, IEEE 802.11, Bluetooth, advanced mobile phone services (AMPS), digital AMPS, global system for mobile communications (GSM), code division multiple access (CDMA), local multi-point distribution systems (LMDS), multi-channel-multi-point distribution systems (MMDS), and/or variations thereof.
Depending on the type of wireless communication system, a wireless communication device, such as a cellular telephone, two-way radio, personal digital assistant (PDA), personal computer (PC), laptop computer, home entertainment equipment, etc., communicates directly or indirectly with other wireless communication devices. For direct communications (also known as point-to-point communications), the participating wireless communication devices tune their receivers and transmitters to the same channel or channels (e.g., one of a plurality of radio frequency (RF) carriers of the wireless communication system) and communicate over that channel(s). For indirect wireless communications, each wireless communication device communicates directly with an associated base station (e.g., for cellular services) and/or an associated access point (e.g., for an in-home or in-building wireless network) via an assigned channel. To complete a communication connection between the wireless communication devices, the associated base stations and/or associated access points communicate with each other directly, via a system controller, via the public switch telephone network, via the Internet, and/or via some other wide area network.
Each wireless communication device includes a built-in radio transceiver (i.e., receiver and transmitter) or is coupled to an associated radio transceiver (e.g., a station for in-home and/or in-building wireless communication networks, RF modem, etc.) that performs analog signal processing tasks as a part of converting data to a radio frequency (RF) signal for transmission and a received RF signal to data.
As is known, the transmitter includes a data modulation stage, one or more intermediate frequency stages, and a power amplifier. The data modulation stage converts raw data into baseband signals in accordance with the particular wireless communication standard. The one or more intermediate frequency stages mix the baseband signals with one or more local oscillations to produce RF signals. The power amplifier amplifies the RF signals prior to transmission via an antenna.
As is also known, the receiver is coupled to the antenna and includes a low noise amplifier, one or more intermediate frequency stages, a filtering stage, and a data recovery stage. The low noise amplifier receives an inbound RF signal via the antenna and amplifies it. The one or more intermediate frequency stages mix the amplified RF signal with one or more local oscillations to convert the amplified RF signal into a baseband signal or an intermediate frequency (IF) signal. As used herein, the term “low IF” refers to both baseband and intermediate frequency signals.
A filtering stage filters the low IF signals to attenuate unwanted out of band signals to produce a filtered signal. The data recovery stage recovers raw data from the filtered signal in accordance with the particular wireless communication standard. Alternate designs being pursued at this time further include direct conversion radios that produce a direct frequency conversion often in a plurality of mixing steps or stages.
As an additional aspect, these designs are being pursued as a part of a drive to continually reduce circuit size and power consumption. Along these lines, such designs are being pursued with CMOS technology thereby presenting problems not addressed by prior art designs. For example, one common design goal is to provide an entire system on a single chip. The drive towards systems-on-chip solutions for wireless applications continues to replace traditionally analog signal processing tasks with digital processing to exploit the continued shrinkage of digital CMOS technology.
One approach of current designs by the applicant and assignee herein is to reduce analog signal processing performance requirements and to compensate for the relaxed performance requirements in the digital domain to provide required system performance. This approach is beneficial in that, in addition to the reduced silicon area requirements, the processing is insensitive to process and temperature variations.
Applications for which this trend is observed include RF receivers where the received signal is digitized as early as possible in the receiver chain using a high dynamic range analog-to-digital converter (ADC), and in a variety of calibration circuits of the radio where signal levels must be measured accurately over a wide range of values. This trend thus increases the demand for embedded low-power, low-voltage ADCs providing high dynamic range in the interface between the analog and digital processing.
A class of ADCs capable of providing high dynamic range and particularly suitable for low-power and low-voltage implementation is known as continuous-time delta sigma analog-to-digital converters (CTΔΣADCs). These ADCs can be designed to operate with supply voltages in the range 1.2V–1.5V and current consumption as low as a few hundred μAs.
FIG. 1 shows an example top-level block diagram of the simplest CTΔΣADC, namely the first-order lowpass CTΔΣADC. The input signal to the CTΔΣADC is a voltage source labeled s(t). An op-amp with negative capacitive feedback constitutes an integrator formed by the operational amplifier and capacitor in a feedback loop, which integrates the input current labeled is(t) flowing from an input signal s(t) to produce an analog integrator output voltage. A coarse (in this example 2-bit) quantizer converts the analog integrator output voltage signal to a digital format shown as y(t). The quantizer, by providing a two-bit output, defines which of four voltage levels most closely match the analog integrator output voltage. More specifically, the quantizer produces a two-bit output having values of 00, 01, 10 and 11.
The quantizer consists of an array of comparators, essentially 1-bit ADCs, whose output is either “high” or “low” depending upon the magnitude of the integrator voltage relative to a reference signal generated by a reference generator. A digital-to-analog converter (DAC) provides a feedback current responsive to a logic value (“1” or “0”) of ADC output to the integrator. FIG. 2 shows one implementation of the 2-bit quantizer and the 2-bit feedback DACs. The quantizer sums the output values of the array of comparators to produce the two-bit output discussed above.
FIG. 3 shows an alternative model of the first-order CTΔΣADC of FIG. 1, wherein the quantizer has been replaced with an additive noise source q(t). The model of FIG. 3 is a model that represents the CTΔΣADC of FIG. 1. Because the operation of the quantizer is deterministic, a signal q(t) may be defined such that the CTΔΣADC of FIG. 3 behaves similarly to the CTΔΣADC of FIG. 1. The digital ADC output, here denoted y(t), can then be written as a sum of two terms, namely a term related to the input signal, ys(t), and a term related to the quantization noise, yq(t), i.e.,y(t)=ys(t)+yq(t).  (1)
By employing feedback around the integrator and quantizer combination, it is possible to suppress the quantization noise component yq(t) in a limited frequency range around DC. Specifically, it can be shown that yq(t) results from q(t) being filtered by a first-order high-pass filter, commonly referred to as the noise transfer function, NTF(s), i.e., in terms of Laplace transforms,Yq(s)=NTF(s)×Q(s).  (2)
Similarly, for a low-frequency input signal s(t), it can be shown that the signal component ys(t) equals the input signal, i.e., in terms of Laplace transforms,Ys(s)=S(s).  (3)
The above properties explain the terminology “lowpass” CTΔΣADC; if s(t) is a low-frequency input signal, the ADC output y(t) closely resembles s(t) when considering only the low-frequency region of y(t), i.e., the ADC “passes” signals of low frequency from analog to digital format without alteration. Furthermore, the lowpass CTΔΣADC of FIG. 1 is of first-order since the single integrator gives rise to a first order high-pass filters; more integrators can, in principle, be added to yield higher order filtering of the quantization noise as is described further below. Generally, an Nth order CTΔΣADC contains N integrators.
Ideally, in equation (2), the quantization noise q(t) is uncorrelated with the input signal s(t) and closely resembles white noise of power Δ2/12, where Δ is the quantizer step size (see FIG. 2) as long as the input signal is limited such that the quantizer operates in the no-overload region. In this case, the two terms that constitute y(t) in equation (1) are uncorrelated, or, equivalently, yq(t) closely resembles white noise, uncorrelated with the input, and filtered by the high-pass filter NTF(s). In this case, since NTF(s) is deterministic, the power of the quantization noise measured over a given signal band-width, fc, of the ADC output y(t) can be determined using standard linear systems analysis as
                              P          n                =                              ∫                          f              =              0                                      f              =              fc                                ⁢                                                    Δ                2                            12                        ⁢                                                                            NTF                  ⁡                                      (                                          ⅇ                      j2πf                                        )                                                                              2                        ⁢                                                  ⁢                                          ⅆ                f                            .                                                          (4)            
For a given known input signal power, Ps, the signal-to-noise ratio (SNR)—a measure of the quality of the analog-to-digital conversion process—can then be calculated a-priori according to
                    SNR        =                                            P              s                                      P              n                                .                                    (        5        )            
Some properties of the ideal CTΔΣADC where q(t) resembles white and random noise follow from (4) and (5). For a given fixed fc, which depends upon the particular application, the SNR depends upon the input as would be expected from a linear system with q(t) contributing constant noise power at the output. In other words, any change of signal power leads to an identical change of SNR in the ADC output; suppose, for example, that the signal power is doubled, e.g., increases by 6 dB, it then follows from (5) that the SNR increases by 6 dB.
Being able to a-priori reliably predict the SNR of the analog-to-digital converted signal, as in equations (4) and (5), is extremely important in almost all applications. Having a-priori knowledge of the SNR delivered by the ADC to within tight tolerances allows system designers to quantify the performance and behavior of the overall system under a variety of different operating conditions. In practice, in order to produce the SNR needed for accurate digital processing of the input signal s(t), a digital filter is used to filter out frequency components above fc in the ADC output signal. As a result of this filtering process, the coarsely quantized output of the CTΔΣADC undergoes a significant increase in bit-resolution.
In practice, however, the above stated assumption that q(t) closely resembles white noise uncorrelated with the input s(t) does not hold true for simple CTΔΣADCs, i.e., for 1st and 2nd order architectures. Especially for DC or low-frequency inputs, the quantization noise is periodic, generating what is commonly referred to as spurious noise, or idle tones. In this case, q(t) is correlated with the input signal s(t), and the frequency spectrum of q(t) contains discrete tones whose frequencies and amplitudes depend upon the specific amplitude and frequency contents of the input. It follows from equation (2) that spurious tones will be observed in the output of the CTΔΣADC with amplitudes and frequencies that are input signal dependent. In practice it is difficult, oftentimes impossible, to exactly predict where in the frequency spectrum the spurious noise appears, and small changes in the input may lead to large changes in the spurious noise. Particularly troublesome, sometimes components of the spurious noise may occur within the passband of the digital filter employed to filter out quantization noise beyond fc, while at other times all spurious components fall beyond fc. This leads to a very undesirable property of the A/D conversion process namely that strong peaks and dips in the power of the in-band portion of yq(t) are observed with strong peaks and dips in the SNR as a result. Unreliable and oftentimes unpredictable behavior of the overall system is an unavoidable result.
As examples of the un-predictable spurious behavior of the quantization noise of the first-order CTΔΣADC of FIG. 1 when driven by DC input, FIGS. 4–6 show the output power spectral density (PSD) of the ADC for the cases of DC inputs of 0.25V, 0.55V, and 0.80V, respectively, vs frequency (in MHz) As may be seen, an output y(t) includes tones, for a 0.25 VDC inputs as shown in FIG. 4, at 3 and 6 MHz. Attenuation of the tones at these frequencies is possible with a notch filter if the input voltage is a constant value and good SNR characteristics may be obtained. FIGS. 5 and 6, however, show that the tones change according to the voltage input level. As may be seen in FIG. 5, ten different tones occur in the 0–6 MHz frequency range for an input voltage of 0.55 VDC. As may be seen in FIG. 6, five tones occur in the 0–6 MHz frequency range for an input voltage of 0.80 VDC. Clearly, the number and frequency of the tones are dependent upon the input voltage level. Thus, a filter cannot easily be made that notches out all of these tones in an environment where the input voltage is not a constant level. Thus, a need exists for a modification, or enhancement, of the basic CTΔΣADC architecture to allow for substantially enhanced linear behavior by randomizing and de-correlating the quantization noise from the input signal without significantly degrading the SNR performance.